Researchers at the University of Maryland's (UMD) Perception and Robotics Group have developed a system that allows a drone to fly through very small and completely unknown gaps using a single camera and onboard processing.
The new system is based on a bee-inspired strategy that yields an 85% success rate.
Unlike other autonomous drone systems, the UMD system does not need to create a three-dimensional map of its environment, or even a model of the gap it is trying to fly through. The drone simply recovers a minimal amount of information that is sufficient to complete the task under consideration.
The system relies on an optical-flow technique in which it takes a picture, moves slightly, and then takes another picture. It identifies similar features in each picture, and due to the parallax phenomenon, the farther-away features behind the gap will appear to have moved less than the closer features around the gap.
From IEEE Spectrum
View Full Article
Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA